How to calculate the ROI of workstation automation — a simple model

A conversation about robotising a workstation usually stalls on one question: when will it pay for itself? The answers tend to come in two flavours, both bad — an enthusiastic "quickly" with no numbers, or a forty-row spreadsheet nobody understands. Yet automation ROI can be calculated on a single page, as long as the model has the right components and honest input data.

Below you get that model: four components of annual savings, the payback formula, one numerical example worked through from start to finish, and a list of the things a spreadsheet does not capture. At the end — the typical mistakes that make calculations drift away from reality.

A simple automation ROI model

The whole model fits into two formulas:

• annual savings = (labour-hour cost × hours removed from the station) + profit from extra throughput + reduction in scrap costs − annual maintenance costs,
• payback period = total investment / annual savings.

The total investment is not just the price of the station. It includes integration with existing machines, grippers and tooling, safety guarding, commissioning, training and a stock of spare parts. In a typical automation project, the items "beyond the robot" can make up a significant share of the budget — leaving them out is the first way to produce a fictitious ROI.

The second rule: take the input data from the station itself, not from plant-wide averages. Measure the real cycle time, the real staffing and the actual scrap rate from recent months. The model is exactly as good as its numbers — and the difference between "the operator runs the machine" and "the operator runs the machine for 60% of the shift and spends the rest packing" can change the result of the calculation by half.

The model line items — what to count and how

Count extra throughput only when the station is a real constraint on the flow — automating a station that waits for material anyway will not add a single piece. How to check this is described in the article the production bottleneck.

A numerical example from start to finish

The example is entirely hypothetical — the numbers were chosen to show the mechanism, not to suggest the prices or results of any specific project.

Assume a machine load/unload station working two shifts. Its robotisation is under consideration.

1. Total investment: PLN 400,000 (robotic cell, gripper, guarding, integration, commissioning, training).
2. Labour: we assume a full labour-hour cost of PLN 55. Automation removes 1.5 FTE from the station (the operator partly stays for supervision and other tasks): 1.5 × 1,700 h = 2,550 h per year. Savings: 2,550 × PLN 55 = PLN 140,250.
3. Extra throughput: the station was the bottleneck; the robot also works through breaks and shift tails, giving, for illustration, 6,000 extra pieces per year at a margin of PLN 5 per piece. Profit: PLN 30,000.
4. Scrap reduction: the scrap rate falls from 2.0% to 0.8% at a volume of 250,000 pcs and a cost per defect of PLN 8: 0.012 × 250,000 × PLN 8 = PLN 24,000.
5. Maintenance costs: service, parts, energy — we assume PLN 20,000 per year.

Annual savings = 140,250 + 30,000 + 24,000 − 20,000 = PLN 174,250.

Payback period = 400,000 / 174,250 ≈ 2.3 years.

In most companies such a result defends the project, as long as the product lives on the market longer than the payback period. It is worth calculating the pessimistic variant right away: without extra throughput and with half the scrap reduction, the savings drop to PLN 132,250 and the payback stretches to about 3 years. If the project holds up in that variant too — the decision is safe. Likewise, check how sensitive the result is to the labour-hour cost and to the hours actually removed, because these are the two items that most often drift away from the original assumptions.

What the simple model does not capture

The spreadsheet counts money, but several real effects stay outside it:

• flexibility — a robotic station changes over differently than a person; with frequent product mix changes this is a cost, with stable production an advantage,
• ergonomics and safety — taking monotonous and strenuous operations away from people reduces injuries and absence, which the model does not directly price,
• staff turnover — strenuous stations are the hardest to staff; automation reduces the cost of recruiting and training new people,
• data quality — an automated station reports cycle times and events, which improves planning,
• ramp-up risk — the first weeks after commissioning rarely run at full output, and this should be reflected in the schedule, not pretended away.

Describe these items qualitatively alongside the calculation. The decision-maker should see both the numbers and what the numbers do not cover. Do not, however, force a monetary value onto these effects — artificially invented amounts weaken the credibility of the whole calculation. One table with hard numbers plus a short list of qualitative effects with commentary works better.

Typical mistakes in calculating automation ROI

The most common mistake is counting the headcount alone. In one direction: "the robot will replace two people" — and then it turns out the operator still has to stay for pallet loading, so in reality half an FTE disappears. In the other direction: leaving out throughput and scrap, which makes a project on the bottleneck look unprofitable when the opposite is true.

The second mistake is understating the investment — a budget for the robot without integration, tooling and training. Third: counting extra pieces the market will not absorb. Fourth: feeding the model with the process in its current, disordered state — automating a mess gives you a faster mess, so before the calculation it is worth going through the steps in the article how to prepare a process for automation. The broader context of the profitability threshold is covered in when does production automation pay off.

The fifth mistake is the horizon. A payback period only makes sense against the life of the product: payback in 2.5 years on a contract that ends in two years means a loss — unless the station can be retooled for the next product and you account for that in the calculation. So next to the payback period, write down the horizon assumption and revisit it whenever the order book changes significantly.

One final note: the model is the same whether you are considering an off-the-shelf robotic cell or a special-purpose machine designed for the process. Only the structure of the investment and the implementation risk differ — which is why both options should be calculated in the same spreadsheet and compared honestly.

Summary

Automation ROI is two formulas and six line items: the full labour-hour cost, the hours actually removed, throughput (only at the bottleneck), scrap, maintenance and the complete investment. Calculate a realistic and a pessimistic variant — if the payback holds up in both, the project is ready for technical discussions. And the factors outside the spreadsheet — flexibility, ergonomics, turnover — describe alongside, instead of squeezing them into the numbers.

Want to check whether your workstation makes economic sense for automation? Describe the process via the contact form — within 48 hours you will receive a preliminary quote and an assessment of which components of the model will matter in your case.